Cover Image

Why Diffusion Models Don't Memorize: The Role of Implicit Dynamical Regularization in Training

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Bonnaire, Tony, Urfin, Raphaël, Biroli, Giulio, Mézard, Marc
Format:	Preprint
Published:	2025
Subjects:	Machine Learning Disordered Systems and Neural Networks
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Topological Exploration of High-Dimensional Empirical Risk Landscapes: general approach, and applications to phase retrieval
by: Maillard, Antoine, et al.
Published: (2026)

The Role of the Time-Dependent Hessian in High-Dimensional Optimization
by: Bonnaire, Tony, et al.
Published: (2024)

Theory of Speciation Transitions in Diffusion Models with General Class Structure
by: Achilli, Beatrice, et al.
Published: (2026)

Kernel Density Estimators in Large Dimensions
by: Biroli, Giulio, et al.
Published: (2024)

Memorization and Generalization in Generative Diffusion under the Manifold Hypothesis
by: Achilli, Beatrice, et al.
Published: (2025)

Biased Generalization in Diffusion Models
by: Garnier-Brun, Jerome, et al.
Published: (2026)

On the role of non-linear latent features in bipartite generative neural networks
by: Bonnaire, Tony, et al.
Published: (2025)

Dynamical Learning in Deep Asymmetric Recurrent Neural Networks
by: Badalotti, Davide, et al.
Published: (2025)

High-Dimensional Analysis of Gradient Flow for Extensive-Width Quadratic Neural Networks
by: Martin, Simon, et al.
Published: (2026)

Cascade of phase transitions in the training of Energy-based models
by: Bachtis, Dimitrios, et al.
Published: (2024)

The Rules-and-Facts Model for Simultaneous Generalization and Memorization in Neural Networks
by: Farné, Gabriele, et al.
Published: (2026)

Dynamical Regimes of Multimodal Diffusion Models
by: Albrychiewicz, Emil, et al.
Published: (2026)

The Exponential Capacity of Dense Associative Memories
by: Lucibello, Carlo, et al.
Published: (2023)

The Interplay of Data Structure and Imbalance in the Learning Dynamics of Diffusion Models
by: Nicoletti, Flavio, et al.
Published: (2026)

Training Dynamics of Nonlinear Contrastive Learning Model in the High Dimensional Limit
by: Meng, Lineghuan, et al.
Published: (2024)

How transformers learn structured data: insights from hierarchical filtering
by: Garnier-Brun, Jerome, et al.
Published: (2024)

Grokking as the Transition from Lazy to Rich Training Dynamics
by: Kumar, Tanishq, et al.
Published: (2023)

Bias in Motion: Theoretical Insights into the Dynamics of Bias in SGD Training
by: Jain, Anchit, et al.
Published: (2024)

$L_0$ Regularization of Field-Aware Factorization Machine through Ising Model
by: Okamoto, Yasuharu
Published: (2024)

Controlled Langevin Dynamics for Sampling of Feedforward Neural Networks Trained with Minibatches
by: Zambon, Alessandro, et al.
Published: (2026)

The Capacity of Modern Hopfield Networks under the Data Manifold Hypothesis
by: Achilli, Beatrice, et al.
Published: (2025)

Why Warmup the Learning Rate? Underlying Mechanisms and Improvements
by: Kalra, Dayal Singh, et al.
Published: (2024)

A Generative Diffusion Model for Amorphous Materials
by: Yang, Kai, et al.
Published: (2025)

Dynamic heterogeneity at the experimental glass transition predicted by transferable machine learning
by: Jung, Gerhard, et al.
Published: (2023)

Analysis of Bootstrap and Subsampling in High-dimensional Regularized Regression
by: Clarté, Lucas, et al.
Published: (2024)

Deep Linear Network Training Dynamics from Random Initialization: Data, Width, Depth, and Hyperparameter Transfer
by: Bordelon, Blake, et al.
Published: (2025)

Emergence of Distortions in High-Dimensional Guided Diffusion Models
by: Ventura, Enrico, et al.
Published: (2026)

Implicit bias produces neural scaling laws in learning curves, from perceptrons to deep networks
by: D'Amico, Francesco, et al.
Published: (2025)

Probing the Latent Hierarchical Structure of Data via Diffusion Models
by: Sclocchi, Antonio, et al.
Published: (2024)

A High Dimensional Statistical Model for Adversarial Training: Geometry and Trade-Offs
by: Tanner, Kasimir, et al.
Published: (2024)

A Dynamical Model of Neural Scaling Laws
by: Bordelon, Blake, et al.
Published: (2024)

Regularization, early-stopping and dreaming: a Hopfield-like setup to address generalization and overfitting
by: Agliari, Elena, et al.
Published: (2023)

Nonreciprocal Spin-Glass Transition and Aging
by: Lorenzana, Giulia Garcia, et al.
Published: (2024)

Generalization Dynamics of Linear Diffusion Models
by: Merger, Claudia, et al.
Published: (2025)

Two-Point Deterministic Equivalence for Stochastic Gradient Dynamics in Linear Models
by: Atanasov, Alexander, et al.
Published: (2025)

On the Geometry of Regularization in Adversarial Training: High-Dimensional Asymptotics and Generalization Bounds
by: Vilucchio, Matteo, et al.
Published: (2024)

Diffusion Operator Geometry of Feedforward Representations
by: Reddy, Kanishka
Published: (2026)

Training neural networks with structured noise improves classification and generalization
by: Benedetti, Marco, et al.
Published: (2023)

Sampling Data with Chains of Forward-Backward Diffusion Steps
by: Kang, Hyunmo, et al.
Published: (2026)

The Training Process of Many Deep Networks Explores the Same Low-Dimensional Manifold
by: Mao, Jialin, et al.
Published: (2023)